Descripción de Hallazgos:

Male-specific RNA coliphages, among the smallest autonomous viruses known, are positive sense, single-strand RNA (ssRNA) phages possessing a genome, 3.8 to 4.2 kb, enclosed within a non-enveloped 26 nm icosahedral-shaped capsid (Buchen-Osmond, 2003). These coliphages belonging to the family Leviviridae were initially grouped primarily

according to their serological properites. The Leviviridae family contain two genera,

Levivirus and Allolevivirus which are further subdivided into four major serogroups, I, II, III,

IV and branched subgroups (a,b,c,d) according to serological cross-reactivity (Sundram et al., 2006). Levivirus are subdivided into genogroups I and II and Allolevivirus are subdivided

into genogroups III and IV. It became apparent as early as the 1970s that the four genogroups of FRNA coliphages had somewhat different fecal source and geographic distributions. In the 1990s the development of genotyping methods based on synthetic oligonucleotide probes (Hsu et al., 1995; Beekwilder et al., 1996) made it possible and convenient to genotype FRNA coliphages and better understand their ecology, their value as fecal and viral indicators, their ability to distinguish human from animal fecal wastes and the impacts of these different waste sources on ambient waters. By employing these methods it became apparent that more information regarding source of fecal pollution could be obtained by comparing full-length genomic sequences from FRNA coliphages collected from various animals and water bodies.

Male-specific coliphages have been suggested as a viral indicator for: (1) fecal contamination (Osawa, 1981; Furuse, 1983), (2) enteric bacterial contamination (Gerba, 1987), (3) enteric viral contamination (Grabow, 2001; Leclerc et al, 2000) and (4) risks of gastro-intestinal illness from recreational water exposures (Colford et al., 2007). FRNA coliphages are almost indistinguishable from most human enteric viruses (Grabow, 2001), occur in higher numbers in sewage and wastewater effluents than viral enteric pathogens (Grabow, 2001), their presence implies the presence of pathogenic viruses (Grabow, 2001), and, in a majority of cases, they display fecal-source specificity (Vinjé et al, 2004; Cole et al, 2003; Furuse, 1987; Schaper et al, 2002; Scott et al, 2002; Stewart, 2002; Long et al, 2005).

The focus of this study was to develop and validate a rapid, genogroup-specific molecular assay for the detection of FRNA coliphages as a potential viral indicator of fecal pollution. Before the molecular assay could be developed, a genetic sequence database was generated representing environmental and prototype FRNA coliphage strains from all four genogroups.

A ssRNA viral assay would need a large (at least 5 strains/genogroup) genetic sequence database. To develop a genetic database, 19 FRNA strains were sequenced and compared to the 11 FRNA full-length sequences available in the National Center for Biotechnology Information (NCBI) genetic database (GenBank) for a total of 30 FRNA strains. FRNA phages were collected from water, sewage, and various animals representative of diverse geographical locations (Table 7.1). The field-collected FRNA strains and

prototype strains were represented by phages isolated from the United States, Japan, Europe, Brazil, Taiwan and/or Mexico (Table 7.1). The majority of groups II (80%) and III (75%)

strains were collected at municipal sewage sources or water bodies with the exception of one group II strain (collected from bird droppings), one group III strain (collected from swine lagoon) and one group III strain was from an unknown source. Two out of 10 group I strains (20%) were collected from sewage, 4 out of 10 group I strains (40%) were collected from ambient waters and/or sentinel organisms (oyster, clam, mussel), one strains’ source was unknown and one strain was collected from a dung-hill. Four out of seven group IV strains (57%) were collected from animal sources (bird, gibbon, calves), two strains (29%) were collected from ambient water sources and one strain was obtained from an infant (Table 7.1).

Phages were sequenced by primer walking. A polyadenylated (Poly-A) tail was added to the 3’ end of purified viral RNA, the poly-A RNA was reverse transcribed with a Poly-T reverse primer; the resulting cDNA was used as a PCR template. A gene-specific forward and a poly-T primer used in the PCR mixture produced an approximate one kilobase (kb)

amplicon. PCR products were gel-purified (GenScript, Piscataway, NJ), cloned (Invitrogen, Carlsbad, CA) and sequenced (Sequetech, Mountain View, CA). This process was repeated until the majority (all sequences except 100-200 bp at the 5’ end) of the genome was obtained. A rapid amplification of cDNA ends (RACE) Smart RACE kit (Clontech, Mountain View, CA) was used to amplify the 5’ portion of the genome.

When full-length genome nucleotide sequences were aligned with the published GenBank strains within their respective genogroup, very similar or identical gene mapping, or Open Reading Frames (ORF), were observed for all four Leviviridae genes indicating the

sequence data generated in this study was valid. Sequence similarity among genogroup I strains ranged from 75.27 - 96.67 % with strain fr forming a separate subgroup (Table 5.2, Fig

5.2). Among group II strains, nucleotide sequence similarity ranged from 83.30 to 93.84% with strains DL10, DL20 and GA having the highest sequence identities (93.43-93.67%) whereas strains T72 and KU1 formed a separate subcluster (Table 5.2, Fig 5.2). Among

Allolevivirus group III, two different subclusters were formed. The first subcluster was

composed of strains VK, HL4-9, BR12, BZ1, TW18 and GenBank strain Qβ having a nucleotide sequence similarity ranging from 91.87-95.69%. The second subcluster formed with GenBank group III strains MX1 and M11 having an 87% nucleotide similarity to each other. The nucleotide similarity of strains between the two group III subclusters ranged from 69.77-71.33% (Table 5.2, Fig 5.2). Group IV Allolevivirus shared sequence identities ranging

from 74.90-95.03 % with the closest identities being 95.03% between strains BR8 and BR1. Amino acid composition was similar among genogroups, further validating the

nucleotide sequences. With the exception of strain fr, amino acid number was consistent in each of the four protein types in group I phages (Table 5.4, Appendix B). The capsid protein of all strains in group I was 130 amino acids in length. Levivirus groups I and II capsid

proteins shared a conserved region consisting of a 10 amino acid, FVLVDNGGTG, consensus sequence. Groups I and II maturation protein shared a consensus region RWLELQ at amino acid (aa) positions number 198-203. The length of the maturation protein of groups III and IV varied from 420 aa to 450 aa (Table 5.4) and possessed a mutual conserved aa region

LWLEFRYGL (Appendix B). The length of the capsid protein was 133 and 132 aa for groups III and IV, respectively, and conserved stretches of amino acids occurred in both groups.

An algorithmic approach was selected to construct phylogenetic trees from the nucleotide sequence and amino acid data (Fig 5.2, Fig 5.4). Nucleotide sequences in the

phylogenetic tree of Levivirus group I strains produced two branches, with 9 strains clustered

as MS2-like and strain fr an individual branch (Fig 5.2). Within group II nucleotide

sequences, strains KU1 and T72 formed one branch and strains DL10, DL20 and GA formed a second branch. Allolevivirus group III nucleotide sequences clustered into a MX1, M11

branch and a second branch with Qβ-like strains BR12, VK, BZ1, HL4-9, TW18 and prototype Qβ. Nucleotide sequence analysis formed three branches in group IV strains as follows: 1) HB-P24, HB-P22 and prototype NL95, 2) BR1, BR8 and prototype SP and 3) prototype FI (Fig 5.2). Individual proteins were clustered into phylogenetic trees. In some cases, phylogenetic protein trees formed more subclusters or branches that the nucleotide trees (Fig 5.4). Genome organization, amino acid conservation and identical or very similar

nucleotide start and stop positions supported the Leviviridae genogroup designation. In

addition, eight nucleotides on the 3' termini clearly distinguish the Allolevivirus,

5' TCCTCCCA 3', from the Levivirus, 5' ACCACCCA 3'.

In addition, two new undescribed Levivirus strains which did not hybridize to

previously designed geno-specific hybridization probes (Vinjé et al., 2004) were sequenced. The two unique FRNA strains were collected from North Carolina and Rhode Island. Full- length genomic sequences from the novel strains were compared to nucleotide and/or amino acid sequences from 10 group I strains (MS2, DL1, DL2, DL13, DL16, ST4, R17, J20, M12, fr) and 5 group II strains (T72, DL10, DL20, GA, KUI). Based on full-length genome

sequences and phylogenetic analyses, these novel strains were placed into a “JS” subcluster of genogroup I. Sequence similarities of the maturation, capsid and lysis proteins of the JS strains were very similar to those of the MS2-like group I strains, sharing 99-100%, 98-100%

and 95-100% sequence similarities, respectively (Table 6.3, Fig 6.1). However, the replicase protein sequences of the JS strains were quite dissimilar to the replicase protein sequences of the MS2-like genogroup I strains, displaying a similarity range of 84-85% and a frame shift resulting from a two nucleotide insertion (Fig 6.5). The resuls of this study provide molecular genetic evidence indicative of recombination in two JS strains of FRNA coliphages. The JS strains provided insight to phage ecology and recombination events in natural FRNA strains.

In this study, analyses of complete genomic sequences from 30 FRNA phages plus two novel strains support the known classification scheme. That is, the Leviviridae consist of

two genera and four distinct genogroups. From this analysis an observation was made that to better define the sub-groupings, it may be more reasonable to assign an association to a specific strain name, i.e., Qβ-like instead of genogroup III, subgroup a; MX1-like instead of genogroup III, subgroup b; and in group IV SP-like instead of genogroup IV, subgroup a; and FI-like instead of genogroup IV, subgroup b. Thus, a recommendation based on these

findings would be to dismiss the alphabetical sub-grouping nomenclature.

Rose et al (1997) designed a one-step reverse transcription polymerase chain reaction (RT-PCR) using a single primer set that detects, but does not differentiate, FRNA coliphages. In this dissertation, a RT-PCR was designed to distinguish the four FRNA coliphage groups (I, II, III, IV) and ultimately, to distinguish human vs animal fecal sources. Primer sets were designed based on the complete genomic sequences of 30 FRNA strains. Genogroup specific RT-PCR primers were designed to conserved sequences from a variety of strains (10 strains from group I, 5 strains from group II, 8 strains from group III and 7 strains from group IV) (Table 7.1, Fig 7.2). Unique amplicon sizes were generated to allow rapid visualization of

each genogroup (Fig 7.1). The traditional one-step RT-PCR was developed, optimized for use with heat-released viral nucleic acid and tested for cross-reactivity. Rigorous validation to ensure lack of cross-reactivity was performed whereby each primer set was tested against 25 environmental and prototype strains. This assay was developed, in part, to allow laboratories lacking real-time equipment the ability to genotype FRNA isolates.

A limitation of molecular detection is that nucleic acid presence or persistence of the phage is detected and is not necessarily representative of the presence of infectious viruses. It has been suggested that if the virion capsid is disrupted, the RNA should degrade rapidly under environmental conditions. This position has not been supported by some lab and field studies which addressed long-term persistence of viral nucleic acids in environmental waters (Kirs andSmith, 2007). Another proposed approach is the discrimination between an intact but non-infectious FRNA phage based on degradation of accessible viral RNA by RNase to eliminate detection of free RNA from damaged (leaky) capsids.

The molecular FRNA phage assay developed in this study may be applicable to an accelerated turnaround time as the assay omits the RNA purification procedure by use of direct heat-release. The traditional primers developed not only allow genogroup identification but provide a comprehensive assessment as to the sanitary quality of the water. If any FRNA phages are detected, then fecal contamination has occurred. This approach utilized a non- cultivation library-independent method for differentiating between human and animal fecal contamination.

In a global phage genotyping assessment, genotypes were reported to be differentially distributed. For example, the FRNA phage from sewage samples in Brazil and West

Germany were from group I exclusively. However, it was unclear as to whether or not sewage treatment plants received slaughterhouse waste (Furuse, 1987). If, however, both slaughterhouse and domestic sewage were combined, and, if only groups II and III are specific to humans, then presumably at least genogroups I, II and III would have been detected in their study. Furuse argued that group I phages observed in raw sewage from treatment plants were most likely introduced from animal sources, and group II and group III phage were from human sources. This begs the question as to why their study only detected group I in sewage treatment facilities from Germany and Brazil. However, several

explanations are possible for this result. Group I could have out-competed groups II and III or they could have slower inactivation rates. This trend of persistent goup I FRNA coliphages should be apparent in other sampling stations if these explanations are correct. However, only limited genotyping from other studies are available for making such observations (Osawa et al.,1981; Miyake et al., 1971). Certain FRNA genogroups may predominate in various human and/or animal populations and their occurrence may also be influenced by climate, diet, intestinal fauna, etc. Further study is needed in different geographical locations to better understand the ecology or natural history of the different FRNA coliphage genogroups; methods developed in the current study should contribute to this effort. Despite the lack of an absolute association between an FRNA genogroup and a unique source, these coliphages likely signal the presence of fecal contamination from either animal and/or human origin. Thus, an FRNA positive sample(s) warrants further investigation, intervention and under some circumstances a public notification alert.

The following three paragraphs will discuss governmental and organizational standards for sanitary quality of recreational water as they pertain to applications of this dissertation. Many of our nation’s ambient water resources are impaired and fail to achieve US EPA implemented water-quality standards. In a report dated April 29, 2008, the number of national impaired waters was 39,918 and the leading impairment was pathogens, totaling 9191 of reported impaired waters, or 14.16% (iaspub.epa.gov/waters/national_rept.control). A water body is defined as impaired when the water body fails to maintain water quality standards even after applying effluent limits for point sources (Clean Water Act Section 303(d); ww.epa.gov/waterscience/standards/303.htm). Water quality standards are based on water quality conditions and pollution sources, i.e., pathogens, nutrients, sediments, metals, habitat alteration and specific chemicals (www.epa.gov/waterscience/standards/about/).

The EPA Beaches Environmental Assessment and Coastal Health Program (BEACH) Act of 2000 was decreed to improve public health and recreational water quality. Fecal contaminated recreational waters may pose a potential health risk as bathers could contract a waterborne disease spread by fecal-oral route

(www.epa.gov/waterscience/beaches/report/chapter02.pdf). Beach-goers who swim or bath in fecal-contaminated waters are at a greater risk than non-swimmers for contracting

gastroenteritis. To protect public health, the use of fecal indicators as water quality standards were implemented (EPA, 1983; EPA, 1984). The BEACH Act (amended Section 303 of the Clean Water Act) requires states, tribes and territories to integrate EPA’s water quality criteria, E. coli and/or enterococci, as their water quality standard (EPA, 2003). The focus of

methods, predict pollution, to better define the criteria as to which fecal indicators and water quality standards are based, to invest in health and methods research and to inform the public (EPA, 2003). The BEACH Act requires states to adopt water quality standards that are “as protective of human health” as the federal criteria. The federal water quality criteria

“Ambient Water Quality Criteria for Bacteria in Recreational Waters, 1986” was developed from US EPA epidemiology studies conducted over a period of years (1972 - 1978) at

beaches located in New York, Louisiana, Massachusetts and Egypt (EPA, 1983; EPA, 1984). Epidemiological data supported the use of E. coli and enterococci as primary fecal indicators

associated with statistically significant increased gastrointestinal illness rates to

swimmers/bathers (EPA, 1983; EPA, 1984). A review of epidemiological studies and fecal indicators (Pruss, 1998) concluded the following: 1) an exposure-response relationship exists in recreational waters between bacterial indicator counts and gastrointestinal symptoms in exposed beach-goers (swimmers) and 2) there was no demonstrated relationship between bacterial indicator counts and non-gastrointestinal symptoms, i.e., rash, eyes, nose, ears.

The primary aim of the World Health Organization’s (WHO) “Guidelines for Safe Recreational Water Environments” (2003) is to protect the public health by addressing such issues as exposure to sewage-contaminated waters, exposure to freeliving pathogenic organisms such as Vibrio, Aeromonas sp. and N. fowleri, exposure to contaminated beach

sand and other potential exposures encountered in recreational water use. Unlike the EPA epidemiology studies conducted in the United States, the WHO based their selection of fecal water quality indicators and health-risk outcomes on a series of randomized control trials conducted in the United Kingdom (REF). The selected bacterial fecal indicator for marine

waters was enterococci whereby a dose-response relationship between enterococci density and health outcome, i.e., gastrointestinal illness and acute febrile respiratory illness (AFRI), was demonstrated. However, the WHO document did not recommend or find a statistical salient

fecal indicator for freshwaters. Enterococci are also the EPA fecal indicator for monitoring marine recreational waters and E. coli is an indicator in freshwaters.

Obstacles to current fecal indicator methods include bacterial culture-based methods, or if a molecular assay is used, the protocols usually involve bacterial or viral RNA/DNA concentration and purification steps thereby increasing the time frame between sample collection, data analysis, public health intervention and protection. Most culture-based detection methods currently require at least a 24-48 hr time lag from sample collection to outcome and therefore, provide information that fecal contamination occurred within the past 24-48 hr. Public health intervention to protect bathers prior to exposure would necessitate sample analysis and data confirmation to occur within hours of sample collection, not days. This lag time between sample collection, completion of analysis and public notification

causes a window of potential risk to bathers from exposure to pathogens. Clearly, real-time or short-term detection, with limited (1-4 hr) turnaround time from water sample collection to results to public notification are imperative for timely protection of bathers.

Although both the EPA and WHO have developed rigorous recreational water quality guidelines, one limitation is that the bacterial indicators have little or no correlation to the presence of pathogenic viruses (Griffin et al., 2003). Bacterial indicators may be an erroneous predictor of viral presence as their survival rates do not match those of viruses. Even when

current bacterial standards are met in recreational waters, risks to human health may be posed by viruses. For example, most illnesses contracted by swimmers appear to be of viral etiology (Griffin et al., 2003). Erroneous bacterial counts have been documented as these fecal

indicator bacteria periodically occurred naturally in temperate climates (Hardina and Fujioka, 1991; Roll and Fujioka, 1997; Byappanahalli and Fujioka, 1998; Fujioka and Byappanahalli, 2000; Solo-Gabriele HM et al., 2000; Genthner et al., 2005). Elevated bacterial indicator counts exceeding EPA water-quality criteria were influenced by soil run-off and not a result of sewage input (Byappanahalli and Fujioka, 2004).

To date, a viral indicator has not been mandated for regulatory purposes in recreational waters. Additional gaps for determining fecal contamination are that these methods are not real-time nor do they provide information regarding source. To minimize risks to human health, resource managers and human health advisors need an early-warning indicator, an indicator that addresses fecal source and an indicator indicative of enteric viral presence.

This attribute as to the selection of a fecal indicator(s) is based on the relationship